CV risk factors and biomarkers are similarly associated with incident HF in men and women
Sex-specific associations of cardiovascular risk factors and biomarkers with incident heart failure
Introduction and methods
Lifetime risk estimates in individuals with heart failure (HF) are comparable in males and females, yet the biological response to HF precursors differs [1-6]. For instance, female hearts tend to remodel in a concentric way after pressure or volume overload, while the hearts of men tend to remodel in an eccentric pattern. The exact mechanism for these sex-related differences in HF remodeling of the heart is unclear, but certain biomarkers that differ in expression levels in men and woman and are linked to pathophysiological processes may provide a better understanding [7-11].
This study evaluated sex-specific association of CV risk factors and biomarkers with the development of HF and to what extent certain biomarkers could improve HF risk prediction in men and woman.
Baseline medical data were harmonized across and pooled from 4 different well-characterized community-based, longitudinal cohorts: the Framingham Heart Study (FHS) offspring cohort examination 6 (1995 to 1998), the Prevention of REnaL AND Vascular Endstage Disease (PREVEND) examination (1997 to 1998), the Multi-Ethnic Study of Atherosclerosis (MESA) examination 1 (2000 to 2002), and the Cardiovascular Health Study (CHS) examination 1 (1989 to 1990; 1992 to 1993). The well-defined primary endpoint in all these studies was the development of HF or death. Exclusion criteria were: <30 years of age, prevalent HF, missing clinical covariates, or unavailable follow-up data. The following biomarkers were included: BNP/NTpro-BNP, cTns (cTnT or cTnI), PAI-1, D-dimer, fibrinogen, CRP, galectin-3, sST2, cystatin-C, and urinary-tot creatinine ratio (UACR). All biomarkers were measured in at least 3 cohorts, except the marker sST2, which was measured in only 2 cohorts. The clinical variables used in sex-specific clinical models were: age, smoking, diabetes mellitus, hypertension, BMI, atrial fibrillation, MI, and the presence of left ventricular hypertrophy/left bundle branch block. Of the 22756 participants included in this study, 12087 (53.1%) were women, and 10669 (46.9%) were men. The median follow-up in women was 12.6 years (11.6-13.6 years) and in men 12.4 years (9.7-13.1 years).
- During the median follow-up, 8.1% (n=989) of women developed HF compared to 10.4% (n=1106) of men, resulting in an HF incidence rate of 7.1 (95% CI:6.6-7.5) per 1000 person-years in women and 9.5 (95% CI:8.9-10.1) per 1000 person-years in men. The overall HF risk was lower in woman than men (HR 0.75, 95% CI:0.68-0.82).
- Left ventricular hypertrophy was more prevalent in woman (P<0.001), while MI and atrial fibrillation (P<0.001) were more common in men as well as hypertension, diabetes mellitus, and history of smoking (P<0.005).
- All clinical risk factors included in the multivariable model were significantly associated with incident HF in woman and men (P<0.001). When tested for sex-specific association, women had a >2 fold increase per 10 years of age in the risk of developing HF (HR 2.07, 95% CI: 189-2.28) compared to a 1.8 fold increased risk in men (HR 1.80, 95% CI:1.67-1.95, Pinteraction=0.001).
- The majority of the individual CV biomarkers were strongly associated with incident HF, but none of the markers showed a significant interaction with sex.
- Natriuretic peptide (NP) levels were strongly associated with incident HF in women (HR 1.47, 95% CI:1.35-1.61) and in men (HR 1.57, 95% CI:1.45-1.70), with no difference in sex (Pinteraction=0.86).
- After further adjustments for NP, there were only 3 biomarkers left that made a significant contribution to incident HF among woman, namely cTns (HR 1.25, 95% CI: 1.17-1.34, P<0.001), CRP (HR 1.14, 95% CI: 1.05-1.24, P<0.001) and UACR (HR 1.27, 95% CI: 1.14-1.41, P≤0.001), while in men, only cTns remained significantly associated with incident HF (HR 1.15, 95% CI: 1.07-1.23, P<0.001).
- Finally, the biomarkers were tested for their risk predictive value in women and men. The traditional risk model gave a C-statistic for women of 0.815 and 0.797 for men. The addition of cTns, or CRP to the model slightly improved risk predictions in women (ΔC-statistic for cTns or CRP is 0.003) and in men the addition of NPs to the model gave a small improvement (ΔC-statistic for NPs is 0.006).
Using data from 4 pooled cohorts to determine sex-specific associations of biomarkers and CV risk factors with incidence HF revealed a strong but similar association among women and men.